Re:What does it mean when an LLM “hallucinates” & why do LLM...
by rcxdude on Tuesday August 22, 2023 @09:42AM
(#37222590)
It's reasonable to call minimising the training loss function the "objective" of the LLM, I think (said loss functions are often called objective functions, after all), though we must careful that the overloading such term in an anthropomorphic fashion. Whether "producing clear and coherent text" is a good characterisation of a "predict the next token" loss function is another question though. I would say it probably isn't, but it's a resonable hypotheses to say that the nature of the training is why the output tends to bias towards plausible but incorrect text as opposed to some variation of "I don't know".
[ Parent
| on HN ]
Re:What does it mean when an LLM “hallucinates” & why do LLM...
by gharman on Tuesday August 22, 2023 @12:55PM
(#37225479)
Agreed there is a semantic argument here. I don’t think anyone is anthropomorphizing LLMs. I’ve had to adopt “objective” with one customer instead of “task” (my preferred default term) because their industry overloads “task” with something else. Now that said, as you point out, the real objective is next token prediction not “help the user with a truth bomb.” and it just happens to have emergent accidental usefulness (I think a natural bias from the training data).
They (chatgpt et al) do move closer to “clear and coherent” by adding extra layers such as a beam search on LLM outputs. Good to remember that ChatGPT et al are products, not bare metal LLMs.
Re:What does it mean when an LLM “hallucinates” & why do LLM...
by chrisjj on Tuesday August 22, 2023 @02:29PM
(#37226751)
> I don’t think anyone is anthropomorphizing LLMs
The "AI" marketeers' use of "hallucinate", "objective" and "predict" does precisely that.
Re:What does it mean when an LLM “hallucinates” & why do LLM...
by chrisjj on Tuesday August 22, 2023 @02:25PM
(#37226704)
> It's reasonable to call minimising the training loss function the "objective" of the LLM
This article isn't doing that, though. What it is calling the model objective is "producing clear and coherent text".
> Whether "producing clear and coherent text" is a good characterisation of a "predict the next token" loss function is another question though.
I agree. But lets us be clear that prediction too is a misnomer. Simply generating a word coherent with prior words is not prediction.